[vtkusers] ICP algorithm question

Wes Turner wes.turner at kitware.com
Thu Jun 14 09:00:13 EDT 2007


Godofredo,

If I am understanding your diagram correctly, the ICP algorithm is doing
what is intended to do.  It minimizes the least squared distance from the
source point set to the target surface.  This necessarily implies
compromises when the source and target do not have the same extent.  Note
that your pink points have been approximately centered over the target
surface, and the position it is in should correspond to the minimum of the
least squared distance.

- Wes

On 6/14/07, Godofredo <godofredoiii at gmail.com> wrote:
>
>
> Hi. I'm making some tries with ICP algorithm but I can't just figure out
> it's
> implementation. Here I show you some images generated with my code. In
> blue
> is the target surface, in green are the source points which I want to
> match
> with the surface and in red are the source points transformed with the
> output of the ICP algorithm. As ou can seee, I've also marked the points
> which best fit beetween two datasets but the ICP algorithm seems to ignore
> the amount of points to fit.
>
> And this are the parameters I use in my code:
>
>         icp->StartByMatchingCentroidsOff();//I use this because I need the
> code for another applications and I       //don't actually want the
> centroids used as source
>         icp->SetMaximumNumberOfIterations(20);
>         icp->SetCheckMeanDistance(true);
>         icp->GetLandmarkTransform()->SetModeToRigidBody();
>
> So this are my questions:How does ICP actually work? why doesn't it fit
> according to the spatialy more closest points beetween the two surfaces?
> Which parameters should I modify and how?
> Thanks in advance.
> http://www.nabble.com/file/p11118686/icp.jpg
> --
> View this message in context:
> http://www.nabble.com/ICP-algorithm-question-tf3921301.html#a11118686
> Sent from the VTK - Users mailing list archive at Nabble.com.
>
> _______________________________________________
> This is the private VTK discussion list.
> Please keep messages on-topic. Check the FAQ at:
> http://www.vtk.org/Wiki/VTK_FAQ
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>



-- 
Wesley D. Turner, Ph.D.
Kitware, Inc.
R&D Engineer
28 Corporate Drive
Clifton Park, NY 12065-8662
Phone: 518-371-3971 x120
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20070614/b711b99f/attachment.htm>


More information about the vtkusers mailing list